Executive Summary - ece4012y202002.ece.gatech.eduece4012y202002.ece.gatech.edu/sd20p02/docs/final... · Web viewThe Portable Bluetooth Stereo Backpack Speaker System was designed

Backpack Speaker System

ECE4012 Senior Design Project

Section L6B, Team BTDESD20P02

Project Advisor, Dr. Robinson

David Rogers - EE - drogers62 @gatech.edu - Team LeaderJack Glennon - EE [email protected]

Andrew Vernetti - CompE - [email protected] - WebmasterAlex Mathis - EE - [email protected]

Connor Brothers - CompE - [email protected]

Submitted

2020 April 29

Executive Summary

The Portable Bluetooth Stereo Backpack Speaker System was designed to be a mobile, durable, and

high quality speaker system. The speakers could connect via Bluetooth to any device or be connected

physically via a 3.5 mm audio jack as input. The final system would consist of a subwoofer box and

two satellite speaker boxes each of which would both contain a woofer and a tweeter. The subwoofer

box would house the circuitry for the system: an amplifier, a microcontroller, and a Bluetooth module,

as well as the power supply system and the circuit boards to contain each subsystem.

As a result of the COVID-19 pandemic, not all of the above original goals were able to be achieved.

Development had just begun on integrating Bluetooth, the amplifier, and the microcontroller when

everything came to a screeching halt. Without the ability to purchase new materials, no development

was able to begin on the physical containment of the speaker boxes. Additionally the power supply

system parts and the high quality speakers were not purchased.

In the end, the working prototype that was developed was the main circuitry behind what would’ve

been a great, high-quality speaker system. The microcontroller communicates with the Bluetooth and

amplifier modules. The speaker system is able to take in an analog input and amplify it to the output

which would be connected to the external speakers. Because no speaker boxes were constructed or

speakers were bought, no analysis can be made of the portability of the system. Additionally, no

analysis can be made on the efficiency relative to the power/battery supply as no power supply exists.

Table of Contents

Executive Summary 1

Table of Contents 3

1. Introduction 4

1.1 Objective 4

1.2 Motivation 5

1.3 Background 6

2. Project Description and Goals 7

3. Technical Specifications & Verification 8

4. Design Approach and Details 11

4.1 Design Approach 12

4.2 Codes and Standards 15

4.3 Constraints, Alternatives, and Tradeoffs 15

5. Final Project Demonstration 17

6. Conclusion 17

7. Leadership Roles 18

8. References 20

Appendix A: QFD Diagram 22

Appendix B: TAS3251 Amplifier Schematic 23

Appendix C: TAS3251 Amplifier 24

1. Introduction

The Portable Bluetooth Stereo Backpack Speaker System was planned to be a mobile, durable, and

quality speaker system that can connect over Bluetooth to any mobile device. The system was designed

to consist of a backpack rig to carry the speakers, a subwoofer, two satellite woofer/tweeter mid-range

speakers, and a mobile app to control the equalization and input of the system, Figure 1 depicts the

original design for the project. Big Type D Energy was awarded $750 to use for the project, but only

$20 was used due to halts in ordering.

Figure 1. Three dimensional rendering of the backpack speaker.

1.1 Objective

The goal of this project was to create a portable backpack system of speakers including a subwoofer to

fill the need of a portable, loud, high quality speaker system. The system was designed to have

sufficient battery life and a user interface that was easy to use. It was meant to provide decent sound

quality and volume, have the equalization controlled through a companion app connected over

Bluetooth, and take audio input via Bluetooth, analog RCA, or 3.5mm jack. Additional features could

have included the ability to connect to up to 4 aftermarket Bluetooth speakers, RGB lighting that

BTDE (ECE4012L6B) Page 4

changes based on the volume and frequency of the music being played, and additional equalization

settings and other signal processing. See Appendix A for a QFD with further details on customer

desires versus engineering requirements.

1.2 Motivation

Currently in the speaker market, there are not many portable speaker systems that are easy to carry

around except for on a cart. Those that are portable are too small to produce meaningful sound pressure

and many have a poor frequency response. An example of a similar style speaker already on the market

is the “Soundboks”. The Soundboks is 216 Watts, 34 lbs, and retails for over $900 [1]. It is battery

powered and transported via handle, but a backpack attachment is made for it and is sold separately for

$129 [2]. In addition to the high cost, the Soundboks does not have tweeters or a dedicated subwoofer

which can lead to both poor highs and lows. It is also a single unit making it unable to produce stereo

imaging. Our speaker system would be more cost effective, higher power, and have the added bonus of

stereo imaging. Another system similar in style is the 1996 Cambridge SoundWorks Model 12. The

Model 12 is 40 Watts, 24 lbs and transportable via suitcase [3]. It has 2 speakers and an amplifier that

are removed from the suitcase which, when closed, doubles as a subwoofer. While this system does

have the capability of imaging, it is unable to fill the need of an outdoor speaker with its small size.

Our backpack speaker system was designed to allow the user to have a decently portable and high

quality sound source for whatever event they would need it. It was also designed to be user-friendly

unlike the Model 12 that requires an amplifier to be wired and unwired with each use. The backpack

speaker system design seeks to fill a hole that the team identified in the speaker market for complete

systems that are portable enough to be carried around and used on battery power but provide high

enough quality to be able to replace normal high-end speaker systems.

1.3 Background


Generally, speaker systems contain 2 full-range speakers and a subwoofer. The purpose for the two

mid-range speakers is to create stereo sound from two different sources of left and right channels

which is how almost all music is recorded. Each of these speakers will typically consist of two

different speaker cones. The first is the woofer which is the main workhorse of the mid-range

frequencies. The second is the tweeter which provides the best frequency response at higher

frequencies. The subwoofer is responsible for the booming bass of the lower frequency notes.

Typically, these speakers would all connect to an external amplifier, where the input from a different

source would come in, be equalized and boosted, then be filtered to the different output speakers

depending on frequency range. In this system, it was instead proposed to have an internal amplification

system that was housed inside of the subwoofer box. Many powered subwoofers have amplifiers in

their boxes already making this a feasible goal [4]. The subwoofer box was designed to take audio

input via a 3.5mm line-in or via Bluetooth. This allows the reduction in complexity of connections to

two speaker cables, one for the left speaker and one for the right with each containing individual wires

for the tweeter and woofer signals. This system would be fully self-contained and would be able to be

used as a standalone system with minimal complexity of parts and subsystems.


2. Project Description and Goals

The team’s design was to implement a portable speaker system consisting of three main components: a

mobile app, mid-range speakers, and a subwoofer. The mobile app would allow, via Bluetooth

connection, the user to control the source input to the speakers, as well as what equalization settings

were desired. Most of the focus of the project was spent on the class D amplifier which turned out to

be quite the challenge to debug and implement properly. One of the biggest goals was worked on was

getting sound to be produced from a small testing speaker.

The original project goals included the following:

General

● Speaker system should be lightweight enough to reasonably be carried on a backpack frame● Total cost less than $500● System should be weather resistant and durable to withstand frequent unloading and carrying● Contain own amplification system with digital crossovers

Mobile App

● Control source input via Bluetooth to the speakers● Choose different equalization settings/digital signal processing functions to run on the

microcontroller inside the speaker system● Send commands to control the volume of the speaker system

Speakers

● Be able to produce high quality sound with low frequency response● Can be powered for a reasonable time from a battery pack

Some of the original goals were not accomplished due to the COVID-19 outbreak leaving group

members unable to meet in person or access campus. These included assembling of the actual speaker

system itself and working towards portable power for it. Communication was able to be established

between the MSP432 and the Bluetooth module, however progress was not made as far as to produce

sound via Bluetooth connectivity.


3. Technical Specifications & Verification

As group members were unable to meet in person, the scope of the project was adjusted to

accommodate the limited access to the project as well as the inability to order new parts. In order to

prove the concept of the original project, it was decided that the audio system that would have been

inside the backpack speaker could still be built or at least prove that the original project could have

been built.

The original specifications are listed in Table 1 and Table 2. The overall project has been severely

limited with most of the original specifications being non-applicable such as final power specifications,

enclosure weight and dimensions, and acoustic specifications. The final power specifications were not

able to be assessed due to the inability to build the entire system. The power consumption of a single

amplifier with an MBED and the Bluetooth module was measured to be 1.68 Watts with the amplifier

out of standby. With the amplifier’s output stage off, the measured power drops to 0.36 Watts.

Feature Specification

Maximum frequency response ripple 0.5 dB

Speaker/box combo alignment Butterworth (B2 or B4)

Power output 200 W

Driver sensitivity (at 2.83 V) >90 dB

Frequency range 20-20,000 Hz

Table 1. Acoustic Specifications

Feature Specification

Total weight <40 lbs

Processing & Bluetooth power .5 W


LED lighting power 1.5 W

Processing delay <100 ms

Setup time <1 min

Startup time <5s

Maximum dimensions 2’ x 1.5’ x 3.5’

Cable length 6’

Auxiliary simultaneous bluetooth devices 1-4

Power supply voltage 24 V

Analog input 2 RCA, 1 3.5 mm

Battery life @ 200 W 1 hrs

Table 2. General Specifications

The final specifications of the project were scoped down to just consisting of the audio system of the

backpack speaker. This included: auxiliary analog input conversion to digital for signal processing

within amp, Bluetooth communication with digital I2S out, amplifier with digital I2S input, and

matched output filters to support 4 Ohm speaker loads. The revised specifications are listed below in

Table 3.

Table 3. Revised specifications for just the system’s electronics.

Processing and Bluetooth power <.5 W

Processing delay <100 ms

Startup time <5s

Auxiliary simultaneous Bluetooth devices 1-4

Power supply voltage 24 V


First, on the communication side of the system, the main audio communication to the amplifier needed

to be PCM over I2S to support high quality audio signals which was accomplished but limited to a

single amplifier due to lack of additional supplies. The original plan was to utilize a high quality ADC,

the TLV320ADC6140, to convert the auxiliary analog input to I2S, however, due to power supply

instability causing voltage spikes on the 3.3V line, the ADC stopped working correctly and the

decision was made to switch prototyping input methods to an MBED in order to have any sort of PCM

over I2S data while the Bluetooth module was still in development resulting in some losses in audio

quality. Additionally, a LaunchPad was used with a Bluetooth receiver since there was a module, the

CC2564MODA, that allowed two-way communication between the system and the streaming device

while also supporting I2S digital out.

Secondly, the enclosure of the system was unable to be assembled. However, the design for the

enclosure should still be able to hold the speakers as well as the audio system while still remaining

portable. The enclosure would have been composed of ½” MDF board with aluminum braces to

provide optimal acoustics for the system while also being able to be flexible with weight. In order to

hold the speakers it was estimated that the midrange cabinet size should be around 0.0983 ft3 and the

subwoofer cabinet size be around 1.06 ft3 in order to have Butterworth alignments for both

cabinet/speaker systems. These dimensions were chosen to hold a 12in subwoofer, 6in midrange

speaker, and a 1in tweeter. This would lead the enclosure size to be between 2-3ft in width and 1-1.5ft

in depth with a height of 3-3.5 ft.

Additionally, in order to power the system it was determined that the 350W MeanWell LRS-350-24

120V AC to 24V DC power supply would be suitable to power the amplifiers’ output stages. The

voltage would need to be stepped down to 12 volts to power the amplifier’s gate drivers and then

further stepped down to 3.3 volts to power all the other components of the system[5]. To achieve the

voltage step down of 24 volts to 12 volts, a switching DC-DC converter, the VXO78012-500-M-TR,


would be used [6]. Tto get the 3.3 voltage a linear regulator, the AZ1117CH-3.3TRG1, would be used

to convert the 12 volts down to 3.3 volts [7].

Figure 2. Power supply Schematic

Finally, in order to get the audio system to turn on and operate, the amplifier had registers that needed

to be initialized in order for the amp to work. The group achieved this through I2C communications

from the MBED which powered up the device, waited for clock initialization, and lastly wrote to all

the desired registers with the appropriate values. Since the MBED was also being used as the ADC

with I2S output, the code for both had to be organized and set up to be isolated as individual functions

that are called as needed. Additionally, when the Bluetooth connection was added, the startup routines

were ported over to the MSP432 LaunchPad.


4. Design Approach and Details

4.1 Design Approach

The first decision that was made was that the entire system should be modular so that if one thing

needed a new revision or failed and needed to be replaced, it would be easier and less time consuming

to do. This led to the Bluetooth module and amplifier boards being developed separately but at the

same time with an agreed upon audio communication standard of PCM over I2S with varying

frequency and a 16-bit minimum word size.

An MBED LPC1768 was used as a host processor for the development of the amp’s setup routine as

well as to output audio streams over I2S to test the digital to analog functions of the amplifier. A

prototype board was fabricated using a CNC mill in order to mount the ADC and amplifier for

breadboard use. This board was then utilized in a large breadboard setup that included all the necessary

capacitors, resistors, and inductors to support the chips. A bench power supply and function generator

were used to generate the necessary voltages and clock signals required. It was left to version two of

the board to include the necessary clocking hardware since it was uncertain what the clocks would

need to be at that time. The amplifier was realized to be the critical path since without it there wouldn’t

be any audio at all and it was proving to be a challenge to get functioning, so most of the development

and debugging was focused on the amplifier / DAC chip.


Figure 3. Block diagram of the speaker system

Several things weren’t able to be resolved, in the end. The signals coming out of the amplifier’s DAC

were extremely noisy when fed data from the MBED’s ADC over I2S. When trying to isolate the

problem by generating a sine wave in software instead of through the ADC, the problem wasn’t able to

be replicated due to the DAC outputting garbage data consisting mostly of spikes which was indicative

of a clocking error experienced earlier in the project. Clock sources for the amp were switched to the

function generator from the sample clock of the I2S bus (as was standard), but although the data

through the digital output of the DSP has less periods of data loss, the DAC still would output garbage

data.

Several Bluetooth modules were considered, but one stood out that was specifically made for playback

of audio, so it was chosen. TI also made a “booster board” for this chip, so a TI LaunchPad was used

as a host processor for configuration and error reporting.


The final stage of the processing and Bluetooth was to combine the two subsystems so that the library

developed on the MBED would operate on the LaunchPad since the LaunchPad could control the

ADC, amp, and Bluetooth module using I2C and UART. Since the LaunchPad and Bluetooth module

had separate PCBs, the amplifier and ADC would have been mounted separately to make a stack of

PCBs four tall to limit the total wire lengths, and keep packaging small and compact.

The speakers for the system were chosen based on their dispersion pattern, sensitivity, required cabinet

volume, weight, and cost in that order. The tweeter chosen was the Dayton Audio NHP23Ti-4 whose

frequency response graph is shown in figure 3. It had a frequency response with a minimum sensitivity

nearing 90 dB at 9 kHz which is the minimum allowed by the project specifications [8]. The midrange

speaker chosen was the Dayton Audio RS150-4 for its smaller cabinet volume requirement while still

remaining above the minimum sensitivity as shown in figure 4 [9]. The subwoofer was purchased

before the start of the project since it was cheap, but its specifications also closely matched the

project’s specifications [10].


Figure 4. Frequency response of Dayton Audio NHP25Ti-4 Tweeter showing the sensitivity never

going below 90dB for the operating range >2kHz (top line) [8].

Figure 5. Frequency response of Dayton Audio RS150-4 midrange woofer showing the sensitivity

never going below 90dB within the operating range of ~100Hz to ~2kHz [9].

4.2 Codes and Standards

The most significant standard that had to be adhered to was the PCM over I2S communication between

the audio chips since that proved to be the most flexible standard. As such, the exact standard (bit rate,

word size) had to be ironed out into something that was common in its implementation throughout the

project. There were also the Bluetooth and I2C communication protocols, but those were more plug

and play as compared to the I2S and didn’t pose much of an issue during the span of the project.


4.3 Constraints, Alternatives, and Tradeoffs

Many different amplifiers exist that could have done almost exactly the same thing as the amplifier that

was chosen, but the chosen TAS3251 was unique in that it integrated a high power amp with a signal

processor and DAC. An alternative amplifier that was considered was the TAS6424 which was a four

channel amplifier that would have been enough to drive the midrange drivers and tweeters but not the

subwoofer. The amount of power necessary for the subwoofer was more than a bridged TAS6424

would be able to handle, so it made sense to only learn one processor / amp instead of two even though

it meant more individual amplifier chips would have to be used.

Another alternative considered was to primarily use analog signal paths instead of the digital ones.

This potentially would have been much easier to diagnose since most of the signals would be exposed

and could be read using an oscilloscope instead of the digital ones that required a lot of interpretation

in order to understand and diagnose. The primary problem with this alternative is that the digital lines

would be stacked on top of the analog lines, posing the problem of cross-talk between the clocks and

communication lines to the pre-amplifier analog signals. This would create a lot of noise in the output

signal from the amplifier and was determined to be too much of a risk. Thus, the signal path was kept

digital.

Much of the project was scrapped since it was no longer practical to interpret the signals in a timely

manner without access to most of the necessary equipment. This also included the soldering and

manufacturing equipment necessary to build the enclosures and to solder various surface mount

packaged components such as the ADC and processors. The prototype boards were used even though

they had severe reliability issues.


5. Final Project Demonstration

The demonstration was difficult to do since many components ended up being fried during the

debugging process when the amplifier’s heatsink was bumped and shorted several communication and

power lines together causing total failure of the amplifier and suspected partial failure of the bluetooth

module, but some demonstrations were done between earlier steps of the project. There were two

primary modules, the bluetooth module and the amplifier module. The bluetooth module had example

code that was modified to support outputting raw I2S data and was tested using an oscilloscope to read

the clocking frequencies and serial data. The amplifier’s power was supplied by a benchtop power

supply that measured output current and voltage while the outputs and inputs of the amplifier were

measured with an oscilloscope to verify functionality. Several things were able to be demonstrated in

videos available on the project website. http://ece4012y202002.ece.gatech.edu/sd20p02/

● show power consumption

● change frequency to adc and see/hear it change on output of amp

● Show bluetooth pairing and audio clock output after transmitting music

6. Conclusion

After the digital circuitry of the signal processor and DAC were found to be more difficult to

program/setup than anticipated, it was considered whether the transition to analog should be made

despite the downside of possible signal degradation. It was then decided to stick with digital since the

encountered problems seemed likely to be overcome within a reasonable amount of time. In hindsight,

it would have greatly simplified the debugging of the project if the processing had been switched to

analog especially with limited ordering capability and debugging equipment.

The current status of the project is that the analog section of the amp works flawlessly with no

perceivable distortion. However, when getting an output from the DAC successfully, most of the time


the output signal is not correct. The programming of the DSP within the amp proved to be extremely

difficult and was actually intended to be programmed by a GUI made by TI using their evaluation

module making it impossible to be utilized. Inputting a sine wave of varying frequency into the ADC

of the MBED and then outputting that onto the I2S bus resulted in a DAC output fed into the amplifier

that produced a waveform where it was easy to distinguish the signal from the noise. This, however,

was with a signal to noise ratio of about 8dB as measured using the FFT function of a Tektronix

TDS3012B oscilloscope.

The MBED resources were difficult to debug as compared to the Bluetooth hardware that was already

setup with well supported libraries, so it was decided to move to debugging the Bluetooth module

instead since it’s part of the final hardware. The Bluetooth module successfully connected to audio

sources, but communication between it and the amp was not established since the amplifier was fried

during the debugging process. The final status of the module integration was trying to get the amplifier

to successfully clock based off of the sample clock provided by the bluetooth module during audio

transmission but the clock was too jittery causing the phase locked loop of the amplifier to fail to lock.

If the project were to be done again, the best way to go about it would be to start with a signal

processor that was better supported by TI and had more examples of its use in DIY projects like

Analog Devices ADAU1701.

7. Leadership Roles

The leadership roles that were needed for the project were Webmaster, Expo Coordinator, Documenter,

Integration Coordinator, and Coding Master. The Webmaster handled the programming and design of the

website, displayed relevant information about the project including technical review papers, accepted

proposal papers, and demonstration videos. The Expo Coordinator was responsible for registering for the

Expo promptly before it was cancelled. The Documenter was responsible for meeting minutes and the final

review of all written documents needed for the class and the project. The Integration Coordinator was


responsible for establishing the compatibility and connection between the components and systems

necessary for the project. Finally, the coding master was in charge of making sure all coding standards are

adhered to, including the organization and functionality of the programs needed in the project.


8. References

[1] The New SOUNDBOKS - The Loudest Portable Bluetooth Performance Speaker, Accessed on:

April, 2020. [Online]. Available: https://www.amazon.com/New-SOUNDBOKS-Bluetooth-

Performance-Swappable/dp/B07WSNJ2WF/ref=sr_1_3?

dchild=1&keywords=soundboks&qid=1587162459&sr=8-3&th=1

[2] Backpack for SOUNDBOKS, Accessed on: April 27, 2020. [Online]. Available:

https://www.amazon.com/SOUNDBOKS-Backpack-for/dp/B07Z9FQS6Q

[3] Instructions for: The Model Twelve Transportable Component Music System by Henry Kloss,

Accessed on: April 27, 2020. [Online]. Available: http://ecx.images-amazon.com/images/I/82dzuc

%2B9a9L.pdf

[4] Subwoofer Plate Amps, Accessed on: April 27, 2020. [Online]. Available:

https://www.daytonaudio.com/category/200/subwoofer-plate-amps

[5] MeanWell LRS-350-24 Power Supply - 350W 24V, Accessed on: April 27, 2020. [Online].

Available:

https://www.amazon.com/MeanWell-LRS-350-24-Power-Supply-350W/dp/B077B36KPB

[6] Non-Isolated DC Switching Regulator, VX078-500-M, Rev. 1.01, CUI Inc, 2020. [Online].

Available: https://www.cui.com/product/resource/vxo78-500-m.pdf

[7] L:ow Dropout Linear Regulator, AZ1117C, Rev. 4-2, Diodes Incorporated, 2019. [Online].

Available: https://www.diodes.com/assets/Datasheets/AZ1117C.pdf


[8] Titanium Dome High Power Neodymium Tweeter 4 Ohm, NHP25Ti-4, Dayton Audio. [Online].

Available: https://www.parts-express.com/pedocs/specs/275-109--dayton-audio-nhp25ti-4-

specification-sheet.pdf

[9] 6" Reference Woofer 4 Ohm, RS1504, Dayton Audio. [Online]. Available: https://www.parts-

express.com/pedocs/specs/295-372-dayton-audio-rs150-4-spec-sheet-revised.pdf

[10] Thunder 5500, T5512-44, MTX Audio. [Online]. Available:

https://www.mtx.com/i/mtxcom/archive/manuals/subs/TechData_T5512-44.pdf


Appendix A: QFD Diagram



Appendix B: TAS3251 Amplifier Schematic


Appendix C: TAS3251 Amplifier


Documents

Executive Summary - ece4012y202002.ece.gatech.eduece4012y202002.ece.gatech.edu/sd20p02/docs/final... · Web viewThe Portable Bluetooth Stereo Backpack Speaker System was designed